Links to HEAD: | (view) (annotate) |
Sticky Revision: | |
Sort logs by: |
In latest commit method insertIdsToKeyMaster wasn't yet complying with the new way of dealing with key Now modified insertIdsToKeyMaster accordingly. Now we create a ClusterConnection object to call the createNewKeyMasterTable method.
Got rid of keyColumn Now key is directly the keyColumn name, e.g. key is asu_id not asu This makes things less confusing and simpler. We've done this following last commit of DataDistribution to also change the ways we deal with key/keyColumn. Note that now the checkKeyCounts and checkDifferingKeys are using to get the keyTable name a ClusterConnection object and the new ClusterConnection method getTableOnNode
New method getIdSetsFromNodes, equivalent to splitIdsIntoSets but rather a posteriori, when tables are already distributed in cluster's nodes Modified method getAllIds4KeyAndTable, now taking host as parameter too, more general Changed the comments a lot for all those 3 methods. There are many catches to be taken in account, the comments explain them quite well
KEYMASTERDB had value for testing, change that back.
Added method insertIdsToKeyMaster Added call to the new insertIdsToKeyMaster in splitTableToCluster method
Added another clusterExecuteQuery method to take a HashMap of queries rather than only one. To be used in cases where the queries to be sent to nodes are different for each node (like in load/dump methods of this class) Right now strictly serial must paralelize it as with the other clusterExecuteQuery
Made MASTER final static variable also public. It is in use by other classes.
Initial commit of DataDistribution.java into aglappe repository Copied from repository scripts-jose rev 69 If want to have a look at previous history must do it from repository scripts-jose, there was no way of copying with history between different repositories
This form allows you to request diffs between any two revisions of this file. For each of the two "sides" of the diff, enter a numeric revision.